Corpus Linguistics

On the Role of Context in Syntax and Semantics
Corpus Linguistics Bibliography

Aim
Example
Context in syntax
Context in the lexicon
Conclusion

Bernhard Kettemann

Aim

There has been a certain tendency to neglect the role of context in syntax and semantics in linguistic teaching and descriptive practice over the past 30 years. Now, the time seems right for a re-assessment, because of the growing awareness that syntax and the lexicon cannot be adequately described without reference to each other, and because British contextualism is getting increasing support from the results of corpus analysis. The aim of this paper is to show that the role of context is important. It is important because the semantics of a word and its syntax are closely related to the choice of other words in its context. In syntax the context selects the particular grammatical form and in semantics the context selects the particular meaning of a specific word. This means that for me there is a close relationship between syntax and semantics, and syntax is not as autonomous as has been assumed so far, especially by members of the generative school, but is limited by lexical choice. But on the other hand, syntax also constrains lexical choice.

Example

An example will help to clarify what I mean: verbs that occur within a similar syntactic frame will have similar meanings. Take the syntactic frame of a verb followed by a prepositional phrase consisting of the preposition by followed by an NP indicating an amount in per cent. Corpus evidence shows that the verbs within this frame belong to a single meaning group: the increase/decrease group (cf. Sinclair/Francis 1996:IV and 173). Let us have a look at the following corpus evidence in form of a concordance. A concordance is a list of occurrences of a particular word, part of a word or combination of words, in its contexts drawn from a text corpus, in this case the MicroConcord Corpus. The search word is sometimes also referred to as keyword. The most common way of displaying a concordance is by a series of lines with the keyword in context (KWIC-format).

Concordance lines 1 to 15 present evidence that our claim that any grammatical structure restricts the lexis that occurs in it is not falsified: the verbs in front of by belong to a single meaning group. On the other hand it is important to point out that such restrictions are not absolute, but more or less clear tendencies: grammar is probabilistic in nature. Note lines 16 to 20: the verb in 16 to 18 is not directly in front of by and still, lifted, boost, reduces belong to the same meaning group, while in lines 19 and 20 the syntax is as desired, but the meanings of the verbs in front of by present a problem in so far as they do not readily fit in with our increase/decrease definition. "Native speakers have only limited intuitions about such statistical tendencies. Grammars based on intuitive data will imply more freedom of combination than is in fact possible" (cf. Stubbs 1996:25-44).
This example has shown that concordancing or corpus analysis is an extremely powerful hypothesis testing device on vast amounts of data. Concordancing allows controlled speculation, makes hidden, but typical structures visible, at the same time enhances imagination and checks it by inductivity, thus making higher degrees of objectivity possible. I argue for the extended use of corpus analysis as another method of gathering data in linguistics, besides observation, elicitation, introspection, and experimentation.
Corpus linguistics provides syntactic and semantic evidence in the sense of what is typical in language use on the basis of frequency, as opposed to intuition and observation. This is an important and in my opinion, crucial `added' value to this type of linguistic analysis.
Corpus linguistics provides additional information on domain specific syntactic and semantic variability in different discourse types, text types and situation types, information that is vital to our ability to understand properly, and that has been largely neglected by mainstream linguistics. This information has been regarded as uninteresting and has been labelled performance. What I hope to show is that this attitude is wrong and harmful to linguistics as a science. Corpus data is not performance, but a systematic record of performance. To illustrate my point I would like to refer to an example by Sir Karl Popper (1994:7). Meteorologists record temperatures all over the world. This is a record of `temperature performance', but it is not temperature, it is a record. Just like corpus data is not language performance, but a record of such performance. Analysing these records allows both, meteorologists and linguists alike, to gain insights beyond the variability of the performance data into the surprising or expected (depending on your theoretical point of view) regularity and systematicity that governs this performance.
My first example has also shown that particular syntactic frames tend to require similar meanings. But this dependency works both ways. Our next example will show that different syntactic frames force a meaning difference on otherwise identical material.

We know that the meaning of the forms of to be to express the notion of future arrangement.
Now compare this meaning with the meaning in the following concordance of to be to in an if-sentence.

Basically I claim (I am grateful to Tim Johns for drawing my attention to these examples, cf. Johns (1997:44-48), and to Georg Marko for discussing them with me) that here the forms of to be to do not express a meaning of future arrangement but instead the modality of desire, while the main clause carries a meaning of necessity, obligation or urgency. I therefore argue that a change in the syntax changes the semantics of an item and a change in the semantics requires changes in the syntax of an item. In short: selection of context is decisive for syntax and semantics. As the context can in principle change indefinitely, it follows that syntax does not consist of a finite set of rules, but an infinite one.

Context in syntax

The distinction between Past Tense and Present Perfect is difficult to grasp for speakers of Southern varieties of German because they do not make this distinction and only use Perfekt, the formal correlate in German of Present Perfect. This means we can expect errors like the following: *Yesterday, I have played with my dog.
A simple concordance of Past Tense uses opposed to Present Perfect uses clearly shows the different conditions on the use of these tenses and could thus be useful as teaching material. It is also interesting to note that the frequency ratio of Past to Present Perfect is about 10:1 in written texts. The reason for this disparity lies in the different overall functions of the two tenses in English. What are these basic functions?

From such concordance evidence it becomes clear for the learners that we have to associate Past Tense with definite past events as opposed to events in the indefinite past with the Present Perfect. The Present Perfect is used to express such notions as Recentness in 1, 2, Result in 3, 4, Duration in 5, 6, 7, 8, and Experience in 9, 10. The Past Tense is used when there is a specific, definite date or point in time given in the context.
The different context dependent functions and uses of much and many often confuse EFL learners. Much is used as an adverb with verbs and as a premodifier of adjectives and as a quantifier determining non-count nouns. Line 1 shows adjectival premodification, in lines 2 and 3 we have the comparative construction as much as, lines 4, 5, 6, 7 and 8 show much in its function as an adverb with verbs, lines 9, 10, 11, and 12 exemplify the quantifier use with non-count nouns:

Many on the other hand is used as a quantifier with count nouns, cf. e.g. lines 1-4. In lines 5 and 6 many functions as a pronoun.

Context in the lexicon

Concordancing can provide many illustrative examples of the semantic relations between words, like homonymy and polysemy, that are disambiguated by context:

Here we find the meanings of race relating to sports in lines 1, 2, and 3, to a metaphorical expansion of the sports meaning to political campaigns in lines 4, 5, and 6, to a metaphorical expansion of the sports meaning into the field of global politics in lines 7, 8, and 9, to people of a particular race in line 10, and to humanity in general in lines 11 to 13.
Metaphorical expansion of meaning is a prime supplier of our word-formation needs. This is possible, because the context provides the information that activates the proper syntactic and semantic feature selection.

The different meanings are: in 1: noun, `good and useful thing, worth copying'; 2 and 3: as in 1, but adjective; 4: `type of car', 5: `types of cars (or other products)', 6: `someone whose job is to show clothes, hair styles etc.', 7: `someone you should imitate, a good person'.

This word is another example of a metaphorical expansion of general English meanings into a business context. In lines 1 to 3 we have the general meaning, in lines 4 to 6 we have a business context, which triggers a different interpretation.

Conclusion

I hope to have shown that context is important for the co-selection of semantic and syntactic structures. This contextual interdependence of syntax and semantics shows that syntax is not independent of the decisions taken on other linguistic levels. Additionally, I hope to have shown that concordancing gives us the possibility to look at performance and offers a window on competence. Its main effect will be a reduction of grammatical and lexical prejudice and linguistic ignorance. This is, I assume what our teaching is about.

References and further reading

Johns, T. (1986), "Micro-Concord, a Language Learner's Research Tool," System 14:2, 151-162.
Johns, T. (1988), "Whence and whither classroom concordancing?" in: T. Bongaerts et al., eds. (1988), Computer Applications in Language Teaching, Dordrecht: Foris.
Johns, T. (1991a), "Should you be persuaded - two samples of data driven learning materials," in: T. Johns/Ph. King, eds. (1991), Classroom Concordancing, ELR-Journal 4, 1-13.
Johns, T. (1991b), "From print out to handout: Grammar and vocabulary teaching in the context of data-driven learning," in: T. Johns/Ph. King, eds. (1991), Classroom Concordancing, ELR-Journal 4, 27-45.
Johns, T. (1997), "If our description of language are to be accurate,... A footnote to Kettemann," TELL/CALL XX, 44-48.
Kettemann, B. (1995a,) "On the Use of Concordancing in ELT," Arbeiten aus Anglistik und Amerikanistik 20:1, 29-41.
Kettemann, B. (1995b), "Concordancing in Vocabulary Teaching," Moderne Sprachen 39:1-2, 35-43.
Kettemann, B. (1995c), "Concordancing in Stylistics Teaching," in: W. Grosser/J. Hogg/K. Hubmayer, eds. (1995), Style: Literary and Non-Literary. Contemporary Trends in Cultural Stylistics, New York: The Edwin Mellen Press, 307-318.
Kettemann, B. (1997a), "Using a Corpus to Evaluate Theories of Child Language Acquisition," in: A. Wichmann/S. Fligelstone/T. McEnery/G. Knowles, eds. (1997), Teaching and Language Corpora, Harlow: Longman, 186-194.
Kettemann, B. (1997b), "Concordancing as input enhancement in ELT," in: B. Lewandowska-Tomaszcyk/P. J. Melia, eds. (1997), Practical Applications in Language Corpora, Lodz, Poland: Lodz UP, 63-73.
Kettemann, B. (1998), "Cognitive Constructivism," in: S. Coelsch-Foisner/W. Görtschacher, eds. (1998), Text and Context. Essays in English and American Studies in Honour of Holger M. Klein, Rheinfelden: Schäuble, 183-195.
Kettemann, B./D. Newby (1993), "If-sentences," English Language Teaching News 19, 65-67.
Murison-Bowie S. (1993), MicroConcord Manual. An Introduction to the Practices and Principles of Concordancing in Language Teaching, Oxford: OUP.
Popper, K., (1994), Knowledge and the Body-Mind Problem, London: Routledge.
Quirk, R./S. Greenbaum/G. Leech/J. Svartvik (1985), A Comprehensive Grammar of the English Language, Harlow: Longman.
Sinclair, J. (1991), Corpus, Concordance, Collocation, Oxford: OUP.
Sinclair, J./G. Francis (1996), Grammar Patterns 1: Verbs, London: Collins/COBUILD.
Stubbs, M. (1996), Text and Corpus Analysis, Oxford: Blackwell.
Tribble C. T./G. Jones (1990), Concordances in the Classroom, Harlow: Longman.
© Gunter Narr: Tübingen 1999
Kettemann, Bernhard (1999),,On the Role of Context in Syntax and Semantics",
in: Kettemann, B. and G. Marko, eds., Crossing Borders: Interdisciplinary Intercultural Interaction, 105-114.

top of page